A Study on Anti-Islanding Detection Algorithms for Grid-Tied Photovoltaic Systems Matlab/Simulink Projects



This study analyzes various anti-islanding (AI) protection relays when the islanding condition of Grid-Tied PV (photovoltaic) System appears at the Point of Common Coupling (PCC) between the PV Solar Power System and the power grid. The main purpose of the study is to determine the performance of several AI prevention schemes in detecting the presence of an island, by monitoring the detection time of the islanding condition through different methods. The devices used to implement the methods include over-current and under-current (OI/UI) relays, over-voltage and under-voltage (OV/UV) relays, over-frequency and under-frequency (OF/UF) relays, rate of change of frequency (ROCOF) and Vector Shift relays. The protection was tested in case of complete disconnection of the PV system from the electric power grid and also in case of various grid faults.





Fig. 1. Simulink model of the 100kW Grid-Connected PV Array



Fig.1: Output results of boost converter

Fig. 2. The output result of dc link voltage (V DC) in VSC

Fig. 3. Id and Iq currents (pu) of VSC Control

Fig. 4. The Voltage between phase A and phase B of VSC

Fig. 5. Simulation result in 20kV measurement point of utility grid.

Fig. 6. The RMS value of voltage in PCC.

Fig. 7. The RMS value of current in PCC.

Fig. 8. The output result of frequency in PCC.



This paper studies and compares different AI detection techniques such as passive AI prevention by standard protection schemes: OI/UI, OV/UV, OF/UF, as well as ROCOF and Vector Shift in case of a 100kW Grid-Connected PV Array. The PV System is completely disconnected from EPS and continues to energize a 20kV utility grid at 50Hz, and respectively various grid faults occurs at 5km away from the PCC of the PV System. The effectiveness of different AI detection algorithms is tested and the impact on network fault conditions and relays behavior during islanding is studied. From the results provided by the performed Matlab/Simulink simulations, it was observed that using traditional relays for islanding detection such as the OC or UV resulted

in significantly better performance in respect to detection time of islanding conditions. The ROCOF and Vector Shift relays have a detection time comparable with frequency relays. However, if the ROCOF threshold is exceeded, the formation of an island is quickly detected. The terminal voltage of PV inverter needs to exceed a certain threshold when the frequency is not stabilized by VSC. The UC relay failed entirely to detect the islanding in both analyzed cases. The effects of unintentional islanding were observed from simulation of transient grid faults on a power distribution network. The protection equipment needs to distinguish between islanding event and grid faults. The Grid-Tied PV System protections should detect the fault and trip before islanding occurs as a result of the opening of the circuit breaker in response to a downstream fault. In order to minimize these effects and to perform according to the. international standards, the AI relays have to be inserted at the points where islanding conditions may occur. The theoretical simulation results are useful to select these points and design the AI protection devices for Grid-Tied PV Systems.



[1] D. Rekioua and E. Matagne, Optimization of Photovoltaic Power Systems, Modelization, Simulation and Control. Springer, 2012.

[2] IEEE Std 1547-2003, Standard for Interconnecting Distributed Resources with Electric Power Systems, IEEE, 2003.

[3] R. Teodorescu, M. Liserre and P. Rodríguez, Grid Converters for Photovoltaic and Wind Power Systems. John Wiley & Sons, Ltd., 2011.

[4] CIGRE Working Group B5.34, “The Impact of Renewable Energy Sources and Distributed Generation on Substation Protection and Automation,” CIGRE, 2010.

[5] IEEE Std 1547.2-2008, IEEE Application Guide for IEEE Std 1547™, IEEE Standard for Interconnecting Distributed Resources with Electric Power Systems, IEEE, 2008

Leave a Reply

Your email address will not be published.